Plots of model fit for all 16 models, for all 3 pgroups

Source

Takes model and data in csv fitforplot16m.csv calculated in optimise_withKandEps.Rmd.

Note on parameters

Predictions generated from the best fitting parameter combination for each model.

All models have epsilon and tau. First eight have kappa (to go with Kindness).

Lesions and order

Models have four ‘components’: Kindness, Selection, Inference, Actual causation.

The full model has all four.

Thereafter they are progressively lesioned.

They are lesioned in the following way:

• First block of eight inc. full: has Kindness, with all other modules lesioned dissociatively (naming convention `noX’ means lesioned to remove the X module.

• Next block of eight: follows the same process in order, this time without Kindness.